這裏簡單總結一下Python操做Oracle數據庫這方面的相關知識。只是簡單的整理一下以前的實驗和筆記。這裏的測試服務器爲CentOS Linux release 7.5. 我的實驗、測試、採集數據的一臺機器。html
1:安裝cx_Oracle
這裏簡單介紹一下cx_Oracle，下面一段內容摘自cx_Oracle的做者（Anthony Tuininga ）的一篇博文，詳情參考https://www.oracle.com/technetwork/cn/topics/tuininga-cx-oracle-086962-zhs.htmlpython
cx_Oracle是一個Python 擴展模塊，經過使用全部數據庫訪問模塊通用的數據庫 API來實現 Oracle 數據庫的查詢和更新。爲使用一些專爲 Oracle 設計的特性，還加入了多個通用數據庫 API 的擴展。cx_Oracle 的開發歷時十多年，涵蓋了大多數須要在 Python 中訪問 Oracle 的客戶的需求。2008 年 12 月，一個新的主要版本解決了早期版本的衆多限制，並增添了對 Python 3.0 和 Oracle 新推出的一些特性的支持。linux
cx_Oracle的官方文檔：https://cx-oracle.readthedocs.io/en/latest/# 官方文檔從架構、特徵方面介紹了cx_Oracle，簡單摘錄以下所示：nginx
Architecture
Python programs call cx_Oracle functions. Internally cx_Oracle dynamically loads Oracle Client libraries to access Oracle Database.git
Fig. 1 cx_Oracle Architecture
cx_Oracle is typically installed from PyPI using pip. The Oracle Client libraries need to be installed separately. The libraries can be obtained from an installation of Oracle Instant Client, from a full Oracle Client installation, or even from an Oracle Database installation (if Python is running on the same machine as the database).github
Features
The cx_Oracle feature highlights are:sql
- Easily installed from PyPI
- Support for Python 2 and 3, and for multiple Oracle Database versions
- Execution of SQL and PL/SQL statements
- Extensive Oracle data type support, including large objects (CLOB and BLOB) and binding of SQL objects
- Connection management, including connection pooling
- Oracle Database High Availability features
- Full use of Oracle Network Service infrastructure, including encrypted network traffic and security features
A complete list of supported features can be seen here.
下面介紹安裝cx_Oracle，cx_Oracle組件有多種安裝方式，實在想不出不用pip安裝的理由，由於它實在太簡單、方便了！shell
pip install cx_Oracle
[root@db-server ~]# pip install cx_Oracle
Collecting cx_Oracle
Downloading https://files.pythonhosted.org/packages/ff/95/6c50d13da95de3e438e7d333dbd7b0946a0981c778e8153c73061b018a33/cx_Oracle-7.2.3-cp36-cp36m-manylinux1_x86_64.whl (731kB)
|████████████████████████████████| 737kB 295kB/s
Installing collected packages: cx-Oracle
Successfully installed cx-Oracle-7.2.3
安裝成功後，測試驗證數據庫
[root@db-server ~]# python
Python 3.6.6 (default, Sep 20 2019, 08:20:38)
[GCC 4.8.5 20150623 (Red Hat 4.8.5-39)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import cx_Oracle
>>>
2：安裝Oracle客戶端
這裏咱們打算安裝Oracle Instant Client，能夠去下面官方網址選擇正確的、所需的版本。本次實驗，我選擇了instantclient-basic-linux.x64-11.2.0.4.0.zipruby
https://www.oracle.com/database/technologies/instant-client/downloads.html
https://www.oracle.com/database/technologies/instant-client/linux-x86-64-downloads.html
這裏下載的Oracle Instant Client Zip，安裝過程簡單以下：
1：解壓安裝包文件到指定目錄
mkdir -p /opt/oracle
cd /opt/oracle/
mv instantclient-basic-linux.x64-11.2.0.4.0.zip /opt/oracle/
unzip instantclient-basic-linux.x64-11.2.0.4.0.zip
2：在操做系統安裝libaio package，不然有可能在運行python腳本時遇到錯誤提示「ImportError: libaio.so.1: cannot open shared object file: No such file or directory」
yum install libaio
注意：這個不是必須的，有些狀況下會遇到這個錯誤。
3：將Instant Client永久添加到運行時連接路徑
sudo sh -c "echo /opt/oracle/instantclient_11_2 > /etc/ld.so.conf.d/oracle-instantclient.conf"
sudo ldconfig #
或者將環境變量LD_LIBRARY_PATH設置爲Instant Client版本的相應目錄。
# echo $LD_LIBRARY_PATH
/usr/local/hadoop/lib/native
在/etc/profile中加入
export LD_LIBRARY_PATH=/opt/oracle/instantclient_11_2:$LD_LIBRARY_PATH
# source /etc/profile
# echo $LD_LIBRARY_PATH
/opt/oracle/instantclient_11_2:/usr/local/hadoop/lib/native
mkdir -p /opt/oracle/instantclient_11_2/network/admin
其它平臺的安裝（或者RPM安裝）， 官方文檔https://oracle.github.io/odpi/doc/installation.html#linux 有詳細介紹，這裏就不多此一舉。 RPM的安裝也很簡單，以下所示
#rpm -ivh oracle-instantclient19.3-basic-19.3.0.0.0-1.x86_64.rpm
Preparing... ########################################### [100%]
1:oracle-instantclient19.########################################### [100%]
須要注意的是，安裝是必須版本一致：cx_Oracle、Oracle Client、Python最好一致，不然就會出現不少問題，例以下面這些（學習、測試中遇到的錯誤）
- 1：cx_Oracle 報錯：cx_Oracle.DatabaseError: DPI-1050: Oracle Client library must be at version 11.2
- 2：cx_Oracle.DatabaseError: DPI-1047: Cannot locate a 64-bit Oracle Client library: "/lib64/libc.so.6: version `GLIBC_2.14' not found (required by /usr/lib/oracle/19.3/client64/lib/libclntsh.so)". See https://oracle.github.io/odpi/doc/installation.html#linux for help
ORACLE的鏈接方式
cx_Oracle鏈接ORACLE數據庫的鏈接字符串（Connection Strings）有多種方式，下面簡單整理一下:
首先，咱們能夠看看cx_Oracle.connect這個函數的參數以下：
cx_Oracle.connect(user=None, password=None, dsn=None, mode=cx_Oracle.DEFAULT_AUTH, handle=0, pool=None, threaded=False, events=False, cclass=None, purity=cx_Oracle.ATTR_PURITY_DEFAULT, newpassword=None, encoding=None, nencoding=None, edition=None, appcontext=[], tag=None, matchanytag=None, shardingkey=[], supershardingkey=[])
參數看起來至關多，眼花繚亂的。不過沒有關係，其實你用到的很少，不少參數基本上能夠選擇性忽略。比較經常使用的基本上有下面幾個：
user 用戶名
password 密碼
dsn dsn
encoding 編碼
Easy Connect Syntax for Connection Strings
dbhost.example.com是機器名，也能夠數據庫服務器IP地址，orclpdb1是service names, 注意，這種方式不能使用早期的SID
import cx_Oracle
# Obtain password string from a user prompt or environment variable
userpwd = ". . ."
connection = cx_Oracle.connect(username, userpwd, "dbhost.example.com/orclpdb1", encoding="UTF-8")
若是數據庫不是默認的端口號，那麼須要在鏈接字符串中指定端口號，以下所示：
connection = cx_Oracle.connect(username, userpwd, "dbhost.example.com:1984/orclpdb1",
encoding="UTF-8")
Oracle Net Connect Descriptor Strings
介紹這種鏈接字符串方式前，咱們必須先了解一下函數makedns,它的功能是建立一個dns對象， 以下所示：
cx_Oracle.makedns
( host
, port
, sid=None
, service_name=None
, region=None
, sharding_key=None
, super_sharding_key=None)
dsn = cx_Oracle.makedsn("dbhost.example.com", 1521, service_name="orclpdb1")
connection = cx_Oracle.connect(username, userpwd, dsn, encoding="UTF-8")
另外，你也能夠用建立相似的鏈接描述符字符串，以下所示：
dsn = """(DESCRIPTION=
(FAILOVER=on)
(ADDRESS_LIST=
(ADDRESS=(PROTOCOL=tcp)(HOST=sales1-svr)(PORT=1521))
(ADDRESS=(PROTOCOL=tcp)(HOST=sales2-svr)(PORT=1521)))
(CONNECT_DATA=(SERVICE_NAME=sales.example.com)))"""
connection = cx_Oracle.connect(username, userpwd, dsn, encoding="UTF-8")
Net Service Names for Connection Strings
這種鏈接字符串方式就不用多介紹了，基本上學習ORACLE的入門的時候，就已經瞭解了這方面的知識點
MyDB =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = xxxxx)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = orclpdb1)
)
)
connection = cx_Oracle.connect(username, userpwd, "MyDB", encoding="UTF-8")
JDBC and Oracle SQL Developer Connection Strings
cx_Oracle鏈接字符串語法與Java JDBC和通用Oracle SQL Developer語法不一樣,通常JDBC的鏈接字符串以下：
jdbc:oracle:thin:@hostname:port/service_name、
cx_Oracle的方式以下：
connection = cx_Oracle.connect("hr", userpwd, "dbhost.example.com:1521/orclpdb1", encoding="UTF-8")
用戶名/密碼@Oracle服務器IP/Oracle的SERVICE_NAME")
db_conn=cx_Oracle.connect('test/test123456@10.20.57.24/GSP')
ORACLE的基本操做
下面簡單介紹一下使用cx_Oracle操做數據庫的一些簡單例子
ORACLE的查詢
簡單查詢
import cx_Oracle
db_conn=cx_Oracle.connect('test/test123456@10.20.57.24/gsp.localdomain')
db_cursor=db_conn.cursor()
sql_cmd='SELECT * FROM TEST.TEST'
db_cursor.execute(sql_cmd)
for row in db_cursor:
print(row)
db_cursor.close()
db_conn.close()
注意事項：SQL語句裏面不能加上分號（;)這類結束符，不然就會報 ORA-00911: invalid character錯誤：
若是上面SQL腳本里面多了一個分號，就會出現下面錯誤：
sql_cmd='SELECT * FROM TEST.TEST;'
[root@MyDB python]# python orace_select.py
Traceback (most recent call last):
File "orace_select.py", line 8, in <module>
db_cursor.execute(sql_cmd)
cx_Oracle.DatabaseError: ORA-00911: invalid character
帶參數查詢
import cx_Oracle
db_conn=cx_Oracle.connect("test", "test123456", "10.20.57.24/gsp.localdomain",encoding="UTF-8")
db_cursor=db_conn.cursor()
sql_cmd='SELECT * FROM TEST WHERE ID=:ID'
sql_p_id={'ID':100}
db_cursor.execute(sql_cmd,sql_p_id)
for row in db_cursor:
print(row)
db_cursor.close()
db_conn.close()
若是要獲取多行記錄，可使用fetchall函數。不用使用
import cx_Oracle
db_conn=cx_Oracle.connect("test", "test123456", "10.20.57.24/gsp.localdomain",encoding="UTF-8")
db_cursor=db_conn.cursor()
sql_cmd='SELECT * FROM TEST WHERE ID=:ID'
sql_p_id={'ID':100}
db_cursor.execute(sql_cmd,sql_p_id)
db_records= db_cursor.fetchall()
print(db_records)
db_cursor.close()
db_conn.close()
ORACLE的DML
其實對於INSERT、DELETE、UPDATE的操做，基本上差不不少。下面簡單簡單舉了一個UPDATE例子，沒有從細節和分類（UPDATE、INSERT、DELETE）展開。
#-*- coding: utf-8 -*-
import cx_Oracle
db_conn=cx_Oracle.connect("test", "test123456", "10.20.57.24/gsp.localdomain",encoding="UTF-8")
db_cursor = db_conn.cursor()
sql_cmd = 'INSERT INTO TEST.TEST(OWNER,OBJECT_ID,OBJECT_NAME)' \
'VALUES(:OWNER, :OBJECT_ID, :OBJECT_NAME)'
db_cursor.execute(sql_cmd, ('TEST', 1, 'KERRY1'))
db_cursor.execute(sql_cmd, ('TEST', 2, 'KERRY2'))
db_conn.commit()
db_cursor.close()
db_conn.close()
注意，若是是插入多行記錄，你使用execute就會報以下錯誤：
cx_Oracle.NotSupportedError: Python value of type tuple not supported.
#-*- coding: utf-8 -*-
import cx_Oracle
db_conn=cx_Oracle.connect("test", "test123456", "10.20.57.24/gsp.localdomain",encoding="UTF-8")
db_cursor = db_conn.cursor()
sql_cmd = 'INSERT INTO TEST.TEST(OWNER,OBJECT_ID,OBJECT_NAME)' \
'VALUES(:1, :2, :3)'
insert_record=[('TEST', 1, 'KERRY1'),('TEST', 2, 'KERRY2'),('TEST', 3, 'KERRY3')]
db_cursor.bindarraysize =3
db_cursor.setinputsizes(30,int, 30)
#db_cursor.execute(sql_cmd, ('TEST', 1, 'KERRY1'))
#db_cursor.execute(sql_cmd, ('TEST', 2, 'KERRY2'))
db_cursor.execute(sql_cmd, insert_record)
db_conn.commit()
db_cursor.close()
db_conn.close()
正確的作法是使用executemany：
#-*- coding: utf-8 -*-
import cx_Oracle
db_conn=cx_Oracle.connect("test", "test123456", "10.20.57.24/gsp.localdomain",encoding="UTF-8")
db_cursor = db_conn.cursor()
sql_cmd = 'INSERT INTO TEST.TEST(OWNER,OBJECT_ID,OBJECT_NAME)' \
'VALUES(:1, :2, :3)'
insert_record=[('TEST', 1, 'KERRY1'),('TEST', 2, 'KERRY2'),('TEST', 3, 'KERRY3')]
db_cursor.bindarraysize =3
db_cursor.setinputsizes(30,int, 30)
db_cursor.executemany(sql_cmd, insert_record)
db_conn.commit()
db_cursor.close()
db_conn.close()
我的在總結的時候，發現官方文檔https://cx-oracle.readthedocs.io/en/latest/user_guide/sql_execution.html裏面的例子實在是太詳細了，這裏徹底沒有必要多此一舉。
參考資料：
https://cx-oracle.readthedocs.io/en/latest/index.html
https://github.com/oracle/python-cx_Oracle/tree/master/samples
https://blogs.oracle.com/oraclemagazine/perform-basic-crud-operations-using-cx-oracle-part-1
https://www.oracle.com/technetwork/articles/dsl/prez-python-queries-101587.html